Notes: Psychology 3020 Feb24
CLASS DISCUSSION
Degrees of Freedom
● The teacher instructs students to be careful about that distinction between the number of
subjects and the number of data points.
● When it's little and it's how many people and one thing in just in this table, it's a number
of data points.
○ How about degrees of freedom within?
● We multiply those other two degrees of freedom together. Two times four.
○ 2x4=8
● K minus one is two five minus one and minus one is for two times four gives me eight.
● And so then again, for the total degrees of freedom, you can either add all those up eight
plus four plus two is 14.
● You can look at the source table and say, we have 15 data points, so degrees of
freedom total should be 15 minus one 14.
● Variance
● Squares, which is just another fancy word for variance, is always found by taking some
of the squares divided by degrees of freedom. So we're going to divide across.
● One thousand ninety-two point thirteen divided by two, I got five, forty-six point seven.
○ 1092÷2=546.7
● Variance for subjects seven, twenty-nine point seven three, divided by four.
● I got one eighty-two point forty-three.
● And variance within two ninety-five, twenty-seven divided by eight.
● I got thirty-six point ninety-eight.
● How do I find the X statistic?
● This is where you want to be really careful about which numbers you're dividing.
● So we're skipping this subject and we're calculating five forty six seven divided by thirty
six ninety eight.
● The statistic is fourteen point seven, seven.
● For completeness, sake, we do it as statistics for the subjects, even though we don't
really do anything with it.
● And that one is going to be mean square subjects divided by means we're here four
point nine three.
● At the table, we use the same process as we did before. It's still between and with the
numerator is degrees of freedom between and the denominator degrees of freedom with
that.
● Degrees of freedom between was two.
● Degrees of freedom within eight.
● I don't know why I put it like that fifteen point three.
● Two times four degrees of freedom, between time with degrees of freedom subjects.
● Critical is 4.4, five nine.
● So we reject the null.
● It is more extreme the by label these critical four point four or five nine.
, Notes: Psychology 3020 Feb24
● My statistics live over here somewhere at fourteen point seventy seven.
● I'm in the regex region, which means I reject the null.
● My results are significant pieceless than 0.05.
● And there's a difference somewhere.
● We don't know, where are you?
● What do we have to do next?
● The limitation of the omnibus, no vote, we can find all sorts of a variance to calculate the
statistic, but then we're stuck here.
● We don't know which conditions are significantly different from which other conditions.
● So what do we need to do next?
● Well, you got it, post-hoc test.
● There's a bunch of different ones.
● The book shows you how to do Tukey's, and this is the exact same format as well.
● It's the exact same formula as the two agencies that are between groups and all that.
● There is a standard error and then I plugged that in to calculate the HST, which is kind of
like a statistical differential to do myself.
● And I compare HST to the critical Q value.
● The cute kitchen table is a lot like the table.
● The only difference is that we have cake along the top.
● So I just use the number of groups, not degrees of freedom for the numerator.
● So Case three.
● Degrees of freedom with them is from my open table of eight, so that's going to be my
critical Q value if HST is bigger than that critical queue.
● Then I can say that there is a difference between the pairs that I'm looking at.
● This does mean I need to do an ancient fee for every pair.
● I have three conditions here, so I have three pairs, one to one in three, two, and three.
● But if I have four conditions, five conditions, I'm going to have to do an agency for every
single possible care to make sure I know where all the possible differences are.
● All right.
● I forgot to put an animation in, so we're just going to walk through this all together.
● Standard error, I'm going to pull mean squares with there from the Inova Table.
● And again, I'm dividing it by and don't make my mistake, we're going to divide it by just a
number of subjects.
● So the standard error is two-point seventy-two.
● I'm going to plug that into the HST for each pair to pair one.
● These are the means from those three groups, remember, she fear had a mean of thirty
four point four midgrade beer was thirty four point six, divided by the standard error.
● The agency is negative point of seven and we look at the absolute value for these right.
● The point of seven is not bigger than 4.0 four so cheap and not significantly different
from each other.
● What about the other pairs cheap and high?
● Thirty four point four was cheap.
● High was fifty two point six.
● Divided by the same standard error of six point six nine six one six nine in favor of the
CLASS DISCUSSION
Degrees of Freedom
● The teacher instructs students to be careful about that distinction between the number of
subjects and the number of data points.
● When it's little and it's how many people and one thing in just in this table, it's a number
of data points.
○ How about degrees of freedom within?
● We multiply those other two degrees of freedom together. Two times four.
○ 2x4=8
● K minus one is two five minus one and minus one is for two times four gives me eight.
● And so then again, for the total degrees of freedom, you can either add all those up eight
plus four plus two is 14.
● You can look at the source table and say, we have 15 data points, so degrees of
freedom total should be 15 minus one 14.
● Variance
● Squares, which is just another fancy word for variance, is always found by taking some
of the squares divided by degrees of freedom. So we're going to divide across.
● One thousand ninety-two point thirteen divided by two, I got five, forty-six point seven.
○ 1092÷2=546.7
● Variance for subjects seven, twenty-nine point seven three, divided by four.
● I got one eighty-two point forty-three.
● And variance within two ninety-five, twenty-seven divided by eight.
● I got thirty-six point ninety-eight.
● How do I find the X statistic?
● This is where you want to be really careful about which numbers you're dividing.
● So we're skipping this subject and we're calculating five forty six seven divided by thirty
six ninety eight.
● The statistic is fourteen point seven, seven.
● For completeness, sake, we do it as statistics for the subjects, even though we don't
really do anything with it.
● And that one is going to be mean square subjects divided by means we're here four
point nine three.
● At the table, we use the same process as we did before. It's still between and with the
numerator is degrees of freedom between and the denominator degrees of freedom with
that.
● Degrees of freedom between was two.
● Degrees of freedom within eight.
● I don't know why I put it like that fifteen point three.
● Two times four degrees of freedom, between time with degrees of freedom subjects.
● Critical is 4.4, five nine.
● So we reject the null.
● It is more extreme the by label these critical four point four or five nine.
, Notes: Psychology 3020 Feb24
● My statistics live over here somewhere at fourteen point seventy seven.
● I'm in the regex region, which means I reject the null.
● My results are significant pieceless than 0.05.
● And there's a difference somewhere.
● We don't know, where are you?
● What do we have to do next?
● The limitation of the omnibus, no vote, we can find all sorts of a variance to calculate the
statistic, but then we're stuck here.
● We don't know which conditions are significantly different from which other conditions.
● So what do we need to do next?
● Well, you got it, post-hoc test.
● There's a bunch of different ones.
● The book shows you how to do Tukey's, and this is the exact same format as well.
● It's the exact same formula as the two agencies that are between groups and all that.
● There is a standard error and then I plugged that in to calculate the HST, which is kind of
like a statistical differential to do myself.
● And I compare HST to the critical Q value.
● The cute kitchen table is a lot like the table.
● The only difference is that we have cake along the top.
● So I just use the number of groups, not degrees of freedom for the numerator.
● So Case three.
● Degrees of freedom with them is from my open table of eight, so that's going to be my
critical Q value if HST is bigger than that critical queue.
● Then I can say that there is a difference between the pairs that I'm looking at.
● This does mean I need to do an ancient fee for every pair.
● I have three conditions here, so I have three pairs, one to one in three, two, and three.
● But if I have four conditions, five conditions, I'm going to have to do an agency for every
single possible care to make sure I know where all the possible differences are.
● All right.
● I forgot to put an animation in, so we're just going to walk through this all together.
● Standard error, I'm going to pull mean squares with there from the Inova Table.
● And again, I'm dividing it by and don't make my mistake, we're going to divide it by just a
number of subjects.
● So the standard error is two-point seventy-two.
● I'm going to plug that into the HST for each pair to pair one.
● These are the means from those three groups, remember, she fear had a mean of thirty
four point four midgrade beer was thirty four point six, divided by the standard error.
● The agency is negative point of seven and we look at the absolute value for these right.
● The point of seven is not bigger than 4.0 four so cheap and not significantly different
from each other.
● What about the other pairs cheap and high?
● Thirty four point four was cheap.
● High was fifty two point six.
● Divided by the same standard error of six point six nine six one six nine in favor of the
